I don't know if anyone else has run into this, but run my Windows machine with a Dvorak keyboard layout. When using the Flirc GUI in windows and the full keyboard layout you get very unexpected behavior. I clicked on a key, for example the period on the QWERTY keyboard in the GUI. For this example I'll use a '.' (period). The GUI asks what remote button to bind to the "period key" (displayed in text even, not just the character). So it gives every appearance of doing what I intend and press the volume up key on my remote. However, after I complete programming I find that the volume up key now returns a 'v', which is the Dvorak value of the same key. Putting the Flirc into a linux machine (also in dvorak) yields the same 'v' character for volume up. I think at some point the GUI must translate through scancodes? That's the only way I can see that "period key" would end up translated into a "v". It just occured to me that the Flirc might ONLY record and replay scancodes. I could probably test this by sticking my Flirc in normal Qwerty machine and see what the results are. I can certainly understand that the GUI can't be expected to adapt for every possible keyboard layout, but it would help keep things straight if what was recorded to the Flirc matched what is reported by the GUI at least in some way. Rather than hard-coding the key names, maybe it could translate back from scancodes to make the "what button to record for <keyname>?" query show the correct current key? I guess it might still have issues if the keyboard layout on the programming machine differed from the layout of the destination machine, but I don't see a way around that other than to possibly detect a non-QWERTY keyboard and warn the user. On a side note... is there any way to query the Flirc for what keys/combinations have been recorded? I occasionally lose my place when reprogramming (I've got two kids) and I end up having to try every key/remote button combination again to make sure I didn't miss one. Thanks for a great product.